Marketing fraud risk analysis method based on flow network

Through the marketing fraud risk analysis method based on streaming network, phishing websites are identified and prevented, and the problem of low monitoring accuracy of dangerous web page recognition in the existing technology is solved, and higher security and a more stable network environment are achieved, and financial fraud is reduced. and the risk of malware spread.

CN118013140BActive Publication Date: 2024-11-15上海栈略数据技术有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410080448.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-19
Publication Date
2024-11-15
Estimated Expiration
2044-01-19

AI Technical Summary

Technical Problem

The existing dangerous web page identification monitoring methods are difficult to achieve higher monitoring accuracy and maintain a more stable and secure Internet environment, and cannot effectively identify and prevent fraud such as phishing websites.

Method used

Using a marketing fraud risk analysis method based on streaming network, we use the method of determining whether the target jump link belongs to a virtual library, obtain the jump path, conduct first-level abnormality evaluation and second-level abnormality verification, and combine risk assessment and domain name matching degree to output risk warnings Information to protect terminal devices.

Benefits of technology

Effectively reduce the infringement of terminal devices by fraud, reduce financial fraud and malware spread, improve network security, enhance pre-identification capabilities for potential threats and vulnerabilities, avoid man-in-the-middle attacks and information leakage, and improve the system's potential threats timely response.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118013140B_ABST
    Figure CN118013140B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of network risk analysis, and in particular to a marketing fraud risk analysis method based on a flow network, comprising: step S1, performing background pre-identification on a jump process from an original web page to a target web page according to a preset plug-in, so as to obtain a target jump path in the jump process; step S2, a cloud server performs a first-level abnormality assessment on the target web page, and determines a first-level assessment result of the first-level abnormality assessment according to the target jump path obtained by the preset plug-in; step S3, a virtual machine determines whether to perform a second-level abnormality check on the target web page according to the first-level abnormality assessment result, and when performing the second-level abnormality check, the cloud server performs a risk assessment on the target web page according to the second-level abnormality check result; step S4, a terminal server receives the risk assessment information returned by the cloud server, and outputs risk warning information before jumping to the target web page based on the original web page according to the risk assessment information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of network risk analysis, and in particular to a marketing fraud risk analysis method based on a flow network. Background Art

[0002] With the rapid development of network technology, individuals and enterprises are using web pages more and more widely, and the risks faced when accessing web pages are also increasing.

[0003] Dangerous web pages, such as phishing web pages, usually refer to websites where criminals use various means to imitate the Uniform Resource Locator (URL) address and page content of real web pages, or take advantage of vulnerabilities in the real web server program to insert dangerous HTML code into certain web pages of the site in order to defraud users of their bank or credit card account numbers, passwords and other private information.

[0004] This type of dangerous web page can cause great property losses to users, so the identification and evaluation of dangerous web pages is particularly important.

[0005] Chinese Patent Publication No.: CN108173814A. A phishing website detection method, terminal device and storage medium are disclosed. The present invention determines the corresponding target link address according to the POST request code in the web page to be detected, grabs the target web page corresponding to the target link address, obtains the jump link address of the target web page, and when the domain names of the target link address and the jump link address are inconsistent, it is determined that the web page to be detected belongs to a phishing website. The computing resources and storage resources consumed are very small, and the engineering implementation is easy. The detection effect is excellent, and the vast majority of phishing websites that steal user information can be detected.

[0006] This shows that the current dangerous web page identification and monitoring methods are difficult to achieve higher monitoring accuracy and maintain a more stable and secure Internet environment. Summary of the invention

[0007] To this end, the present invention provides a marketing fraud risk analysis method based on a flow network to overcome the problem that the current dangerous web page identification and monitoring methods in the prior art are difficult to achieve higher monitoring accuracy and maintain a more stable and secure Internet environment.

[0008] To achieve the above object, the present invention provides a marketing fraud risk analysis method based on a flow network, comprising:

[0009] Step S1, determining whether the target jump link belongs to the virtual library, if so, performing background pre-identification on the jump process corresponding to the target jump link, and obtaining the target jump path in the jump process;

[0010] Step S2, when the first-level abnormality evaluation is determined to be turned on according to the number of nodes in the target jump path and the maximum threshold of the safe jump node, determine whether the process jump link URL corresponding to each target jump node in the target jump path belongs to the safe library, and count the number of process jump link URLs that do not belong to the safe library. If the number is not 0, calculate the first-level abnormality score of the target jump path;

[0011] Step S3, when it is determined to enable a secondary abnormality check for the target web page reached through the target jump path according to the primary abnormality score and the minimum risk score threshold, a secondary abnormality check is performed on the target web page according to the secondary jump range and the secondary jump allowed safety range of the target web page to obtain risk assessment information;

[0012] Step S4: when there is a target keyword in the risk assessment information, before jumping from the original web page to the target web page, output risk warning information.

[0013] Furthermore, the safe library stores the safe jump link URLs of each safe web page and the safe jump paths of each safe web page; the dangerous library stores the dangerous jump link URLs of each dangerous web page and the dangerous jump paths of each dangerous web page;

[0014] Determining the target jump link according to each safe jump link URL and each dangerous jump link URL;

[0015] If the target jump link does not belong to the safe library and does not belong to the dangerous library, the target jump link belongs to the virtual library.

[0016] Further, determining to start the first-level abnormality assessment according to the number of nodes in the target jump path and the maximum threshold of the safe jump node includes:

[0017] Obtaining the number of target jump nodes in the target jump path;

[0018] Generating a safety jump node fluctuation curve according to the number of safety jump nodes of each safety web page in the safety library, and determining the maximum threshold of the safety jump node of the safety page according to the safety jump node fluctuation curve;

[0019] If the number of the target jump nodes does not exceed the maximum threshold of the safe jump nodes, it is determined to start the first-level abnormality assessment.

[0020] Further, for any target jump node in the target jump path, the process jump link URL corresponding to the target jump node is identified;

[0021] If the process jump link URL does not belong to the secure jump link URL stored in the secure library, it is determined that the process jump link URL meets a single determination condition.

[0022] Further, the first-level anomaly score K is calculated according to the acquired monitoring data;

[0023]

[0024] Among them, H is the first-level abnormality compensation parameter. For the i-th jump process from the i-th target jump node to the i+1-th target jump node in the target jump path, the traffic peak before and after the jump, the interaction frequency Pi with the terminal server during the jump, and the jump duration Ti are monitored, and the traffic used for this jump Qi is determined based on the traffic peak before and after the jump.

[0025] Further, the first-level abnormality compensation parameter H=Y+a1×a2;

[0026] Among them, a1 is the network status evaluation parameter of the terminal server during the jump from the original web page to the target web page, a2 is the web design evaluation parameter of the target web page, and Y is the number of jump link URLs in each process in the target jump path that meet the single judgment condition.

[0027] Further, the minimum risk score threshold is determined according to the historical monitoring data between each risk jump node of each risk page in the risk library;

[0028] If the first-level anomaly score does not exceed the minimum risk score threshold, it is determined to start the second-level anomaly check.

[0029] Further, the risk level of the target web page is determined according to the matching degree between the target domain name of the target web page and each storage repository; wherein the storage repository includes: the safe repository, the dangerous repository and the virtual repository;

[0030] If there is no domain name with a high degree of matching with the target domain name in both the safety database and the danger database, the risk level of the target web page is determined to be level I danger;

[0031] If the target domain name has a high matching degree with any domain name in the risk database, the risk level of the target web page is determined to be level II risk;

[0032] If the target domain name has a high matching degree with any domain name in the security library, the risk level of the target web page is determined to be level III danger.

[0033] Further, determining the allowed safe range of the secondary jump of the target web page at different risk levels according to the risk level;

[0034] The size of the secondary jump allowed safety range is related to the level of the risk grade. The higher the risk grade, the smaller the secondary jump allowed safety range.

[0035] Further, the domain name prefix overlap degree is determined according to the domain name prefixes of each secondary jump webpage domain name that has secondary jumps in the target webpage; and the secondary jump range that has secondary jumps in the target webpage is determined according to the domain name prefix overlap degree;

[0036] If the secondary jump range does not exceed the secondary jump allowed safety range, determining that the target web page is the safe page;

[0037] If the secondary jump range exceeds the secondary jump allowable safety range, the target web page is determined to be the dangerous page, and the risk warning information is output.

[0038] Compared with the prior art, the beneficial effect of the present invention is that the present invention can effectively protect the terminal from fraud by outputting risk warning information before the original web page jumps to the target web page. Dangerous web pages usually impersonate legitimate websites with the purpose of defrauding users of sensitive information, such as user names, passwords, credit card information, etc. Early shielding can reduce the chance of terminals mistakenly entering these malicious websites, thereby reducing the risk of users being attacked and reducing the occurrence of financial fraud, protecting the financial security of users. Dangerous web pages may lead to the spread of malicious software, thereby threatening the entire network security. Early shielding can slow down or curb these attacks, which helps maintain the security of the entire network.

[0039] In particular, by judging the target jump link in the original web page, if it belongs to neither the safe library nor the dangerous library, the preset plug-in is started to pre-identify the jump process of the target jump link, and further risk judgment is made for web pages without storage records to reduce the risk of web page login. Pre-identification helps to discover potential threats and vulnerabilities in advance. By pre-identifying the target web page, security can be enhanced to prevent potential network attacks, data leakage and other security risks.

[0040] In particular, by pre-identifying the jump process from the original web page to the target web page, the target jump path in the jump process is obtained, and by analyzing the number of target jump nodes in the target jump path, it is determined whether the target web page is a dangerous page or needs further abnormality evaluation. Analyzing the web page security based on the number of jump nodes in the path can avoid the risk of man-in-the-middle attacks and enhance data integrity. Fewer nodes reduce the possibility of data tampering and reduce the risk of information leakage. A shorter jump path may reduce the risk of sensitive information leakage. Each jump node is a potential point where information may be leaked. Therefore, the shorter the path, the smaller the chance of leakage.

[0041] In particular, by conducting a preliminary analysis based on the number of target jumps in the target jump path, the jump link URLs of each process of each target jump node in the target jump path are analyzed, thereby avoiding the limitation that only considering the number of nodes in the jump path cannot comprehensively evaluate the security of the web page.

[0042] In particular, by setting different secondary jump allowable safety ranges for target web pages of different risk levels, this differentiated evaluation standard can more finely respond to threats of different degrees and improve overall security; different levels of risks may require different evaluation standards because they may involve different threats and consequences. Customized evaluation standards can more accurately evaluate various risks and thus respond to them more effectively, and can concentrate resources on high-risk websites, so that security teams and system resources can be better allocated to ensure a stronger response in high-risk situations. Customized evaluation standards can reduce the false alarm rate, that is, better distinguish between real threats and false alarms. For low-risk websites, looser standards can be adopted to reduce false alarms for normal websites, thereby reducing unnecessary interference. For high-risk websites, a faster response may be required. By setting different evaluation standards, it can ensure that high-risk events are detected and responded to more quickly, improving the system's timely response to potential threats.

[0043] In particular, abnormal analysis of redirected web pages can help identify and respond to potential security threats and problems. Abnormal analysis can help identify phishing websites, which often try to disguise themselves as legitimate websites to steal users' sensitive information. By monitoring abnormal behavior in the redirect path, possible phishing attacks can be discovered early and help discover possible security vulnerabilities in the website, including detecting unauthorized redirects, insecure transmission protocols and other problems. By intercepting and tampering with data transmission in the communication path, man-in-the-middle attacks and abnormal redirects or information modifications can be prevented. By detecting abnormal redirect behavior, it helps prevent malicious redirects and avoid redirecting users to malicious sites, thereby improving user privacy protection, avoiding situations that lead to user privacy leaks, and improving the ability to identify and respond to various security threats, which helps maintain the security of users and websites. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a flow chart of the marketing fraud risk analysis method based on the flow network in this embodiment;

[0045] Figure 2 This is a logic diagram for determining to enable the first-level abnormality assessment in the marketing fraud risk analysis method based on the flow network in this embodiment;

[0046] Figure 3 This is a logic diagram for determining risk levels in the marketing fraud risk analysis method based on flow network in this embodiment;

[0047] Figure 4 This is a logic diagram for determining to enable secondary anomaly checking in the marketing fraud risk analysis method based on flow network in this embodiment. DETAILED DESCRIPTION

[0048] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0049] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0050] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0051] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0052] See also Figure 1-Figure 4 As shown, Figure 1 This is a flow chart of the marketing fraud risk analysis method based on the flow network in this embodiment; Figure 2 This is a logic diagram for determining to enable the first-level abnormality assessment in the marketing fraud risk analysis method based on the flow network in this embodiment; Figure 3 This is a logic diagram for determining risk levels in the marketing fraud risk analysis method based on flow network in this embodiment; Figure 4 This is a logic diagram for determining to enable secondary anomaly checking in the marketing fraud risk analysis method based on flow network in this embodiment.

[0053] The present invention provides a marketing fraud risk analysis method based on a flow network, comprising:

[0054] Step S1, the cloud server determines the storage library for the target jump link, and when it is determined that the storage library is a virtual library, the preset plug-in performs background pre-identification on the jump process from the original web page to the target web page, and obtains the target jump path in the jump process;

[0055] Step S2, the cloud server determines whether to start the first-level abnormality judgment according to the obtained target jump path, and determines the first-level abnormality judgment result according to the number of jump nodes in the target jump path that meet the single judgment condition;

[0056] Step S3, the virtual machine determines whether to perform a second-level abnormality check on the target web page according to the first-level abnormality judgment result, and when performing the second-level abnormality check, the cloud server performs a risk assessment on the target web page according to the second-level abnormality check result;

[0057] In step S4, the terminal server receives the risk assessment information returned by the cloud server, and outputs risk warning information before jumping to the target web page based on the original web page according to the risk assessment information.

[0058] In the embodiment of the present invention, by outputting risk warning information before the original web page jumps to the target web page, the terminal can be effectively protected from fraud. Dangerous web pages usually impersonate legitimate websites with the purpose of defrauding users of sensitive information, such as user names, passwords, credit card information, etc. Early shielding can reduce the chance of terminals mistakenly entering these malicious websites, thereby reducing the risk of users being attacked and reducing the occurrence of financial fraud, protecting the financial security of users. Dangerous web pages may lead to the spread of malicious software, thereby threatening the entire network security. Early shielding can slow down or contain these attacks, which helps maintain the security of the entire network.

[0059] Specifically, in this embodiment, when the terminal clicks on the target jump link in the original web page, the cloud server determines the target jump link;

[0060] If the target jump link belongs to the security library, then the target web page is determined to be a security page;

[0061] If the target jump link belongs to the dangerous library, the target web page is determined to be a dangerous page;

[0062] If the target jump link does not belong to the safe library or the dangerous library, the cloud server stores it in the virtual library and starts the preset plug-in to perform the pre-identification on it;

[0063] The security library stores the secure jump link URLs of secure web pages and the secure jump paths of each secure web page;

[0064] The dangerous database stores dangerous jump link URLs of dangerous web pages and dangerous jump paths of each dangerous web page.

[0065] The embodiment of the present invention determines the target jump link in the original web page, and starts a preset plug-in to pre-identify the jump process of the target jump link when it belongs neither to the safe library nor to the dangerous library, and further determines the risk of web pages without storage records, thereby reducing the risk of web page login, and the pre-identification helps to discover potential threats and vulnerabilities in advance. By pre-identifying the target web page, security can be enhanced to prevent potential network attacks, data leakage and other security risks.

[0066] Specifically, the preset plug-in in this embodiment performs the pre-identification on the jump process from the original web page to the target web page. During the pre-identification process, the preset plug-in can obtain the target jump path during the jump process.

[0067] The cloud server identifies the number of security jump nodes of each security webpage according to the security jump path of each security webpage in the security library, generates a security jump node fluctuation curve according to the number of security jump nodes of each security webpage, identifies the fluctuation range of the security jump node according to the generated security jump node fluctuation curve, and calculates the average value of the number of security jump nodes according to the security jump node fluctuation curve, thereby obtaining the maximum threshold value X0 of the security jump node of the security webpage;

[0068] If the number X of target jump nodes in the target jump path does not exceed the maximum threshold X0 of safe jump nodes, the cloud server determines to start the first-level abnormality assessment;

[0069] If the number X of target jump nodes in the target jump path exceeds the maximum threshold X0 of safe jump nodes, the cloud server determines that the target web page is a dangerous page.

[0070] The embodiment of the present invention pre-identifies the jump process from the original web page to the target web page, thereby obtaining the target jump path in the jump process, and analyzes the number of target jump nodes in the target jump path to determine whether the target web page is a dangerous page or needs further abnormality evaluation. The web page security is analyzed according to the number of jump nodes in the path, which can avoid the risk of man-in-the-middle attacks and enhance data integrity. Fewer nodes reduce the possibility of data tampering and reduce the risk of information leakage. A shorter jump path may reduce the risk of sensitive information leakage. Each jump node is a potential point where information may be leaked. Therefore, the shorter the path, the smaller the chance of leakage.

[0071] Specifically, in this embodiment, when the cloud server determines to turn on the first-level abnormality assessment, if the target jump path obtained by the preset plug-in is |A1-A1-A3-A4-A5|, where A1 is the first target jump node, A2 is the second target jump node, A3 is the third target jump node, A4 is the fourth target jump node, and A5 is the fifth target jump node, for any target jump node Ai, i=1,2,3,4,5, the cloud server identifies the i-th process jump link URL corresponding to the target jump node Ai;

[0072] If the i-th process jump link URL is the secure jump link URL stored in the secure library, the cloud server determines that the process jump link URL does not meet the single determination condition;

[0073] If the i-th process jump link URL is not the secure jump link URL stored in the secure library, the cloud server determines that the process jump link URL meets a single determination condition;

[0074] For the acquired target jump path |A1-A1-A3-A4-A5|, when the number of process jump link URLs corresponding to the target jump nodes in the target jump path that meet the single judgment condition is 0, the cloud server determines that the target web page corresponding to the target jump path |A1-A1-A3-A4-A5| is a safe page; when the number of process jump link URLs corresponding to the target jump nodes in the target jump path that meet the single judgment condition is not 0, the cloud server determines to calculate the first-level anomaly score.

[0075] The embodiment of the present invention analyzes the URLs of each process jump link of each target jump node in the target jump path based on a preliminary analysis of the number of target jumps in the target jump path, thereby avoiding the limitation that only considering the number of nodes in the jump path cannot comprehensively evaluate the security of the web page.

[0076] Specifically, in this embodiment, when the cloud server calculates the first-level anomaly score, the preset plug-in monitors the jump process in real time, and obtains the traffic peak before and after the jump, the interaction frequency Pi with the terminal server during the jump, and the jump duration Ti during the i-th jump from the i-th target jump node to the i+1-th target jump node, and determines the jump usage traffic Qi according to the traffic peak before and after the jump;

[0077] The cloud server calculates the first-level anomaly score K according to the monitoring data of the preset plug-in;

[0078]

[0079] Among them, H is the first-level abnormality compensation parameter.

[0080] Specifically, the cloud server in this embodiment determines the first-level abnormality compensation parameter H according to the network status evaluation parameter a1 of the terminal server, the web design evaluation parameter a2 of the target web page, and the number Y of the target jump path |A1-A1-A3-A4-A5| that meets the single judgment condition during the jump from the original web page to the target web page, where H=Y+a1×a2;

[0081] The cloud server identifies the number of dangerous jump nodes of each dangerous web page according to the dangerous jump path of each dangerous web page in the dangerous database, and calculates the abnormal score of the jump process according to the obtained interaction frequency, jump duration and jump usage flow during the jump between any two dangerous nodes, and averages the abnormal scores between the dangerous jump nodes in the dangerous database to obtain the minimum danger score threshold;

[0082] If the first-level anomaly score K exceeds the minimum risk score threshold, the cloud server determines that the target web page is a risky page;

[0083] If the first-level anomaly score K does not exceed the minimum risk score threshold, the cloud server determines to start the second-level anomaly check.

[0084] Specifically, in this embodiment, when the secondary abnormality check is turned on, the virtual machine determines whether the target web page is a dangerous web page according to the secondary jump range of the secondary jump occurring in the target web page;

[0085] The cloud server determines the risk level of the target web page according to the target domain name of the target web page;

[0086] Determine the domain name with the highest matching degree in the security library and the risk library according to the target domain name;

[0087] If there is no domain name with a high degree of matching with the target domain name in both the safety database and the danger database, the cloud server determines that the risk level of the target web page is level I danger;

[0088] If the target domain name has a high matching degree with any domain name in the risk database, the cloud server determines that the risk level of the target web page is level II risk;

[0089] If the target domain name has a high matching degree with any domain name in the security library, the cloud server determines that the risk level of the target web page is level III danger.

[0090] Specifically, the virtual machine in this embodiment determines the allowed safe range of secondary jump of the target web page under different risk levels according to the risk level of the target web page determined by the cloud server;

[0091] The size of the secondary jump allowed safety range is related to the level of the risk grade. The higher the risk grade, the smaller the secondary jump allowed safety range.

[0092] In the embodiment of the present invention, different secondary jump allowed safety ranges are set for target web pages of different risk levels. Such differentiated evaluation criteria can more finely respond to threats of different degrees and improve overall security. Different levels of risks may require different evaluation criteria because they may involve different threats and consequences. Customized evaluation criteria can more accurately evaluate various risks, thereby more effectively responding to them, and can concentrate resources on high-risk websites, so that security teams and system resources can be better allocated to ensure a stronger response in high-risk situations. Customized evaluation criteria can reduce the false alarm rate, that is, better distinguish between real threats and false alarms. For low-risk websites, looser standards can be adopted to reduce false alarms for normal websites, thereby reducing unnecessary interference. For high-risk websites, a faster response may be required. By setting different evaluation criteria, it can be ensured that the detection and response to high-risk events are faster, thereby improving the system's timely response to potential threats.

[0093] Specifically, the cloud server in this embodiment obtains the domain name prefix of each secondary jump webpage domain name where the secondary jump occurs in the target webpage; and determines the domain name prefix coincidence degree according to the domain name prefix of each secondary jump webpage domain name and the domain name prefix of the target webpage, including: the first domain name prefix coincidence degree, the second domain name prefix coincidence degree...the nth domain name prefix coincidence degree;

[0094] The cloud server averages the overlap of each domain name prefix, thereby determining the secondary jump range where the secondary jump occurs in the target web page;

[0095] If the secondary jump range exceeds the secondary jump allowed safety range, the virtual machine determines that the target web page is a dangerous page;

[0096] If the secondary jump range does not exceed the secondary jump allowed safety range, the virtual machine determines that the target web page is a safe page.

[0097] The embodiments of the present invention can help identify and respond to potential security threats and problems by performing abnormal analysis on jump web pages. Abnormal analysis can help identify phishing websites, which usually try to disguise themselves as legitimate websites to steal users' sensitive information. By monitoring abnormal behavior in the jump path, possible phishing attacks can be discovered early and help discover possible security vulnerabilities in the website, including detecting unauthorized jumps, insecure transmission protocols and other problems. By intercepting and tampering with data transmission in the communication path, man-in-the-middle attacks and abnormal jumps or information modifications can be prevented. By detecting abnormal jump behavior, malicious redirections can be prevented and users can be avoided from being redirected to malicious sites, thereby improving user privacy protection and avoiding situations that lead to user privacy leaks. The ability to identify and respond to various security threats is improved, which helps maintain the security of users and websites.

[0098] The calculation compensation parameters and the calculation adjustment parameters described in the present invention have two functions: one is to balance the left and right dimensions of the formula, and the other is to adjust the numerical results. No specific assignment is performed in this embodiment. In addition, the calculation formulas in this embodiment are used to intuitively reflect the adjustment relationship between the various numerical values, such as positive correlation and negative correlation. Unless otherwise specified, the parameter values ​​that are not specifically limited to numerical values ​​are all positive.

[0099] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0100] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A marketing fraud risk analysis method based on a flow network, characterized in that: include, Step S1, determining whether the target jump link belongs to the virtual library, if so, performing background pre-identification on the jump process corresponding to the target jump link, and obtaining the target jump path in the jump process; Step S2, when the first-level abnormality evaluation is determined to be turned on according to the number of nodes in the target jump path and the maximum threshold of the safe jump node, determine whether the process jump link URL corresponding to each target jump node in the target jump path belongs to the safe library, and count the number of process jump link URLs that do not belong to the safe library. If the number is not 0, calculate the first-level abnormality score of the target jump path; Step S3, when it is determined to enable a secondary abnormality check for the target web page reached through the target jump path according to the primary abnormality score and the minimum risk score threshold, a secondary abnormality check is performed on the target web page according to the secondary jump range and the secondary jump allowed safety range of the target web page to obtain risk assessment information; Step S4, when the target keyword exists in the risk assessment information, before jumping from the original web page to the target web page, outputting risk warning information; Determining the risk level of the target web page according to the matching degree between the target domain name of the target web page and each storage repository; wherein the storage repository includes: the safe repository, the dangerous repository and the virtual repository; The safe library stores the safe jump link URLs of each safe web page and the safe jump paths of each safe web page; the dangerous library stores the dangerous jump link URLs of each dangerous web page and the dangerous jump paths of each dangerous web page; Determining the target jump link according to each safe jump link URL and each dangerous jump link URL; If the target jump link does not belong to the safe library and does not belong to the dangerous library, the target jump link belongs to the virtual library; Determining to start the first-level abnormality assessment according to the number of nodes in the target jump path and the maximum threshold of the safe jump node includes: Obtaining the number of target jump nodes in the target jump path; Generating a safety jump node fluctuation curve according to the number of safety jump nodes of each safety webpage in the safety library, and determining the maximum threshold of the safety jump node of the safety webpage according to the safety jump node fluctuation curve; If the number of the target jump nodes does not exceed the maximum threshold of the safe jump nodes, it is determined to start the first-level abnormality assessment; Determine the minimum risk score threshold according to the historical monitoring data between each risk jump node of each risk page in the risk library; If the first-level anomaly score does not exceed the minimum risk score threshold, it is determined to start the second-level anomaly check; If there is no domain name with a high degree of matching with the target domain name in both the safety database and the danger database, the risk level of the target web page is determined to be level I danger; If the target domain name has a high matching degree with any domain name in the risk database, the risk level of the target web page is determined to be level II risk; If the target domain name has a high matching degree with any domain name in the security library, the risk level of the target web page is determined to be level III danger; Determining, according to the risk level, the allowed safety range of the secondary jump of the target web page at different risk levels; The size of the secondary jump allowed safety range is related to the level of the risk level. The higher the risk level, the smaller the secondary jump allowed safety range. For any target jump node in the target jump path, identifying the process jump link URL corresponding to the target jump node; If the process jump link URL does not belong to the secure jump link URL stored in the secure library, then determining that the process jump link URL meets a single determination condition; Calculating the first-level anomaly score K according to the acquired monitoring data; ; Wherein, H is a first-level abnormal compensation parameter. For the i-th jump process from the i-th target jump node to the i+1-th target jump node in the target jump path, the traffic peak before and after the jump, the interaction frequency Pi with the terminal server during the jump, and the jump duration Ti are monitored, and the traffic used for the jump Qi is determined according to the traffic peak before and after the jump; The first-level abnormality compensation parameter H=Y+a1×a2; Among them, a1 is the network status evaluation parameter of the terminal server during the jump from the original web page to the target web page, a2 is the web design evaluation parameter of the target web page, and Y is the number of jump link URLs in each process in the target jump path that meet the single judgment condition.

2. The method for analyzing marketing fraud risks based on flow network according to claim 1, characterized in that: Determine the domain name prefix overlap degree according to the domain name prefixes of each secondary jump web page domain name that has secondary jumps in the target web page; and determining the secondary jump range of the target web page where the secondary jump occurs according to the domain name prefix overlap; If the secondary jump range does not exceed the secondary jump allowed safety range, determining that the target web page is the safe web page; If the secondary jump range exceeds the secondary jump allowable safety range, the target web page is determined to be the dangerous page, and the risk warning information is output.